dibend/US-State-Zip-Code-3D-Correlation-Matrix

1import gradio as gr2import pandas as pd3import plotly.graph_objects as go4import numpy as np5 6def plot_real_estate_correlation(state):7    # Read the CSV file8    df = pd.read_csv('https://files.zillowstatic.com/research/public_csvs/zhvi/Zip_zhvi_uc_sfrcondo_tier_0.33_0.67_sm_sa_month.csv')9 10    # Filter for the given state11    df = df[df['State'] == state.upper()]12 13    # Extract the list of ZIP codes and filter only columns that are date strings14    zip_codes = df['RegionName'].unique()15    16    # Extract columns that are valid date strings only17    date_columns = []18    for col in df.columns[7:]:19        try:20            # Try to parse column names as dates21            pd.to_datetime(col)22            date_columns.append(col)23        except:24            continue25    26    # Initialize a DataFrame to hold price data for correlation calculation27    price_matrix = []28 29    # Loop through each ZIP code in the state30    for zip_code in zip_codes:31        df_zip = df[df['RegionName'] == zip_code]32 33        # Extract only the columns with valid date data (price values)34        prices = df_zip.loc[:, date_columns].values.flatten()35 36        # Append prices to the matrix if there are no missing values37        if not np.isnan(prices).all():38            price_matrix.append(prices)39 40    # Convert to DataFrame for easier manipulation41    price_matrix_df = pd.DataFrame(price_matrix, index=zip_codes, columns=date_columns)42 43    # Transpose to align for correlation calculation (each column = ZIP code)44    price_matrix_df = price_matrix_df.T.dropna()45 46    # Calculate the correlation matrix for ZIP codes47    corr_matrix = price_matrix_df.corr()48 49    # Prepare the grid data for 3D plot50    z_data = corr_matrix.values51    x_data, y_data = np.meshgrid(zip_codes, zip_codes)52 53    # Create the 3D surface plot54    fig = go.Figure(data=[go.Surface(z=z_data, x=x_data, y=y_data)])55 56    # Update plot layout57    fig.update_layout(58        title=f'3D Correlation Matrix of Housing Prices in {state}',59        scene=dict(60            xaxis_title='ZIP Code',61            yaxis_title='ZIP Code',62            zaxis_title='Correlation',63        ),64        autosize=True65    )66 67    return fig68 69iface = gr.Interface(fn=plot_real_estate_correlation,70                     inputs=[gr.components.Textbox(label="State (e.g., 'NJ' for New Jersey)")],71                     outputs=gr.Plot())72 73iface.launch(share=False, debug=True)